Advertisement. The page you requested will display in seconds.
Advertisement. The page you requested will display in seconds.
*Ideally, cameras should reproduce the colors of the real world as accurately as possible. Inaccurate colors can make landscapes look unnatural or people look sickly. We test color accuracy by photographing a GretagMacbeth Colorchecker test chart and comparing the colors the camera reproduces with the known colors of the test chart. The ColorChecker consists of 24 tiles of colors from around the color spectrum. The image below shows how accurate the Fuji A920’s colors are compared to the actual colors of the Colorchecker chart. The outside squares show the colors the camera reproduces and the inner squares show the actual chart colors corrected for exposure. The inner rectangles show the actual chart colors at a perfectly even exposure. The photos are taken with the white balance set to Tungsten, which provides the most accurate color reproduction under our tungsten studio lights.
Looking at the chart above, you can see a lot of the color tiles are reproduced quite accurately, although there are a few notable exceptions. The blues are quite inaccurate and look fairly undersaturated, while several other tiles in the third row fare the same. The third row consists of highly saturated colors, which shows the A920 has trouble reproducing such vivid color. The graph below shows color accuracy in a more quantitative way. The squares show the locations of the ColorChecker colors on the color spectrum, and the circles show where the colors the A920 reproduces lie. The lines connecting the squares and circles show the amount of color error for each color tile.
The graph confirms many of the highly saturated colors (the ones farthest from the white circle at the center) are reproduced inaccurately. This is most apparent in the blues and yellows, which has a strong effect on blue skies and green foliage, making them look unnatural. Overall, the A920 has decent color accuracy, though not as good as we’ve seen in some similarly-priced models from other manufacturers.
*Resolution performance in digital cameras has a number of factors other than simply how many megapixels a camera has. Sensor design can significantly affect resolution performance, as can the processor. Cameras will over- or undersharpen photos automatically, possibly making them look sharper but adding ugly image artifacts. We test resolution performance by photographing an industry-standard resolution test chart and varying the focal length and exposure settings. We run these photos through Imatest, which determines sharpness in terms of line widths per picture height (lw/ph), as well as the amount of sharpening applied in-camera. The units lw/ph represent the number of equally-spaced, alternating black and white lines that can fit across the picture frame before becoming blurred.
The 9-megapixel A920 proved to be sharpest at ISO 100, f/4.1, and a focal length of 17.5mm. The camera resolves 1997 lw/ph horizontally with 3.9 percent oversharpening, and 2143 lw/ph vertically with 4.4 percent oversharpening. These numbers are fantastic, and reiterate the impressive resolution Fuji has achieved in its cameras over the past few years, even in budget models like the A920. The amount of sharpening applied is just enough to make the images look slightly sharper without creating ugly image artifacts. These impressive numbers don’t tell the whole story, however. When shooting at longer focal lengths, the camera strongly undersharpens photos, negating its impressive resolution capabilities. More importantly, only the centers of photos are crisply in focus, with the corners and edges looking washed out and blurry. Chromatic aberration is very apparent on the sides of images, and some distortion is evident. Keep these issues in mind if you plan to view or print your photos large, or crop and enlarge, because the flaws will become much more apparent.
Noise – Manual ISO* (7.36)
*"Noise" refers to the graininess digital images show, often in low light or shadows. Digital noise is quite different from film grain, however, and is almost always unwanted. Noise can come in the form of white sandiness or colored splotches, or anything in between. We test noise levels by photographing our test chart under bright, even studio lights at all ISO speeds available. We run the images through Imatest, which measures noise levels by the percent of image detail it obscures.
The A920 keeps noise impressively low throughout its limited ISO range. The noise itself is quite ugly, composed of sharp-edged colored splotches, but is only apparent when looked at very closely. Overall, the A920 receives a very good manual noise score.
Noise – Auto ISO* (1.93)
*We also test noise levels with the camera set to Auto ISO, under the same bright studio lights as above. Under these bright lights, the A920 chose ISO 400, which is a bit puzzling, but didn’t produce enough noise to be very noticeable.
**Still Life Sequences **
Click the thumbnails to view the high resolution images.
White Balance*** (5.05)
*White balance has a big effect on color accuracy in photos. Different types of light have different color casts, from indoor fluorescent to outdoor shade, and cameras must be able to adjust appropriately. We test white balance accuracy by photographing the ColorChecker chart under four types of light: flash, fluorescent, outdoor shade, and tungsten. We test the Auto white balance setting, as well as the appropriate presets. The A920 could not focus on the test chart without a light source shining on it, so we could not confidently evaluate the accuracy using only the flash.
*Auto (3.27) *
With the white balance set to Auto, the A920 is mediocre under white fluorescent light and very poor under outdoor shade and indoor tungsten. The images below show what kind of color cast your photos will have when shooting under these light sources. It is not always the best idea to leave this camera on Auto white balance.
*The presets fare a bit better than the Auto setting. Accuracy is decent under fluorescent light (using the "Fluorescent Light-3" setting), poor in outdoor shade, and excellent under tungsten lighting. It pays to use the white balance presets on the Fuji A920, especially when shooting indoors.
Still Life Sequences
Click the thumbnails to view high-resolution images.
|Still Life Scene|
|ISO 100||ISO 100|
|ISO 200||ISO 200|
|ISO 400||ISO 400|
|ISO 800||ISO 800|
Low Light* (4.18)
*We’ve seen how the A920 performs under bright studio lights, but how do photos look in less-than-ideal shooting conditions? We test color and noise performance in low light by photographing the ColorChecker at 60, 30, 15, and 5 lux. Sixty lux corresponds to a room lit softly by two lamps, 30 lux approximates the amount of light in a room lit solely by a 40-watt bulb, and 15 and 5 lux are extremely dark, testing the limits of the sensor. All shots are taken at the highest ISO speed available, which is ISO 800 on the A920.
The A920 cannot expose properly at 15 or 5 lux, showing the camera has clear limits and cannot capture photos in dim situations where a flash isn’t desired. At 60 and 30 lux the camera’s colors suffer, though noise levels stay relatively low. Overall, this is not a good camera for low light shooting.
We also test long exposure performance, but only at ISO 400 in order to compare cameras evenly. The A920 can only take exposures longer than 0.25 seconds in Night mode (up to 4 seconds), where ISO cannot be set manually, so we could not test long exposure performance. Four seconds doesn’t allow much room to play around with slow shutter speeds, anyway.
Dynamic Range* (5.66)
*Dynamic range is an image quality factor that tells how large a tonal range a camera can detect. This proves important in scenes with high contrast, where a camera must be able to show detail in the bright highlights as well as the dark shadows. These situations show up often at weddings or landscapes shot in bright sunlight. We test dynamic range by photographing a backlit Stouffer step chart at all ISO speeds. The Stouffer chart consists of a long row of gray rectangles that vary in tone from brightest white to darkest black. The more rectangles a camera can discern, the better its dynamic range.
The A920 has good dynamic range at ISO 100, and then drops off slowly but steadily at higher sensitivities. The camera does a good job keeping dynamic range up at high ISO speeds, which is most likely due to its fairly low noise levels. Overall, the camera has slightly better than average dynamic range, which is great for a budget camera.
Speed/Timing – All speed tests were conducted using a Kingston Ultimate 120X 2GB SD Card, with the camera set to highest resolution and best quality, unless otherwise noted.
*Startup to First Shot (6.8)
*The A920 takes a leisurely 3.2 seconds to turn on and fire its first shot.
*Very disappointingly, the A920 does not have a Continuous Shot mode. Capturing action shots will become a shot in the dark with this camera.
*The camera has no measurable lag when the shutter is held halfway down and prefocused, and a lag of 0.4 seconds when not prefocused.
*The A920 takes 3.1 seconds to process one 4.5 MB full-resolution fine-quality shot taken at ISO 200.
Video Performance* (1.04)
Bright Light – 3000 lux *
We record footage of our color charts under bright studio lights set to 3000 lux to see how the video handles color and noise. Under such bright tungsten lights, the A920’s video has terrible color accuracy, which is actually normal for digital camera video. Noise levels, on the other hand, are very low.
*Low Light – 30 lux
*Color accuracy is a little better in low light, but not nearly as accurate as a still photo. The camera cannot properly expose at 30 lux, and noise levels are very high. Forget about capturing low-light footage of your friends in a nightclub or your family at sunset.
*We record footage of our resolution chart to see how the Movie mode handles resolution. The tiny 320 x 240 image frame captures only 302 lw/ph horizontally with -2.7 percent undersharpening, and 253 lw/ph vertically with -5.4 percent undersharpening. Despite the tiny size and lack of oversharpening, imaging artifacts are still apparent.
*We also take cameras outside to see how the Movie mode renders moving cars and pedestrians. As we mentioned above, the frame size of the A920’s video is tiny, and makes it hard to see on a computer screen. The motion has tons of moiré, the exposure "flashes" whenever it changes, and bright highlights bleed across the entire frame. The good news is the exposure usually looks quite accurate, and the frame is so small it is hard to detect any jerkiness in the motion.
News and Features
The K-3 finally gets the makeover it deserves.
These lenses come with a neat trick—OLED displays
Weather-sealed and compact, this Fuji is ready to grab landscapes.
Sony brings advanced features to two affordable CyberShots.
Wearable action cams? 4K pro camcorder? GH4 firmware? Yes!
Already solid point-and-shoots get an update.
Getting up close and personal with our little friends in the backyard.
Chef Jacques la Merde presents junk food in artisanal fashion.
With two new cameras, Canon's 4K lineup is more versatile than ever.
Sign up to get the latest news and reviews only available to our email subscribers
Thank you for subscribing!